Energy Aims to Launch America’s First Exascale Supercomputer by 2021

Denis Starostin/Shutterstock.com

Featured eBooks

The Government's Artificial Intelligence Reality
What’s Next for Federal Customer Experience
Cloud Smarter

Insiders say the first-of-its-kind machine will accelerate innovation and discovery across the health, defense and climate sectors—and beyond.

The Energy Department and Intel plan to deliver the United States government’s first exascale supercomputer by 2021, “dramatically” advancing scientific research and discovery, stakeholders announced Monday.

In a $500 million partnership with Intel and subcontractor Cray Inc., Energy will develop Aurora, a first-of-its-kind supercomputer that will enable high-performance computing, or HPC, and artificial intelligence at exascale. The system will be capable of at least one exaflop, which is a quintillion (that’s a billion billion) calculations per second. The partnership was the result of a competitive procurement process and the system will be delivered at Argonne National Laboratory in Chicago.

“Achieving exascale is imperative, not only to better the scientific community, but also to better the lives of everyday Americans,” Energy Secretary Rick Perry said in a statement. “Aurora and the next generation of exascale supercomputers will apply HPC and AI technologies to areas such as cancer research, climate modeling, and veterans’ health treatments. The innovative advancements that will be made with exascale will have an incredibly significant impact on our society.”

The supercomputer will drive unprecedented innovation by accelerating the convergence of HPC, data analytics and AI at exascale, government and industry leaders told reporters.

Rick Stevens, Argonne National Laboratory’s associate director for computing, environment and life sciences, said the system will be a strong platform for traditional HPC applications, it’s designed to rapidly perform data analytics for streaming data crucial to the Energy (such as data coming off of telescopes, accelerators and detectors), and it will be explicitly configured to be an excellent platform for deep learning.

“Some of the projects I am personally looking forward to on this system are accelerating our efforts and collaboration with the National Institutes of Health and Cancer and with the [Veterans Affairs Department] on such projects as suicide prevention, cardiac risk and also traumatic brain injuries,” Stevens said.

The machine will also have significant defense applications, Paul Dabbar, Energy’s Under Secretary for Science, said. For example, high-performance computing will be utilized to test, analyze and accelerate understanding around strategic weapons. The machine will also improve wind turbine efficiency and experts’ ability to predict climate at a regional scale.

“The 60,000 people who work in the 17 national labs need to have the greatest capabilities to drive energy, science and defense analysis and research and, importantly, also help drive discovery science across the board,” Dabbar said. “The national lab complex here in the United States is the largest generator of Nobel prizes in the world and utilizing the right infrastructure, and in this case high-performance computers, is paramount towards us driving discovery science and all of our other applications going forward.”

Dabbar noted that, last year, the Energy National Lab Complex commissioned the “number one and two supercomputers in the world” and that it also houses five of the top ten supercomputers in the world. While it’s expected to put America on the leading edge, it’s still unclear if Aurora will be the fastest supercomputer in the world when launched in 2021.

“We don’t know what everybody else is doing, so we can really only talk to the plans of the United States,” Stevens said. “We know other countries are working on the path to exascale, but we don’t know precisely when they will deploy their systems."